The multilingual services market has received a lot of buzz through Computational Linguistics (or Natural Language Processing), a multidisciplinary field involving artificial intelligence, computer science and linguistics that uses computational processes to handle human language. Research has provided important applications for the work of translators, such as search engines, spell checkers and voice recognition, as well as memory systems and machine translation. Corpus Linguistics, an area of study within Computational Linguistics, has also had a major influence on translation, including machine translation. This is an area that studies language in use, that is, it investigates language by observing large amounts of authentic data contained in the corpus. To do this, it makes extensive use of computational tools to organize, extract and interpret information from the corpus. Corpus Linguistics considers language as a probabilistic system, i.e. there are many possibilities of expression in language, but not all are used with the same frequency, as some have become conventionalized through use.
The findings of Corpus Linguistics have contributed to conventional and machine translation in various ways. By knowing and employing the standards most commonly used in a language, the result is a translation that is more natural and faithful to the native language. In addition, most machine systems use a system based on a corpus consisting of bilingual texts (original and translated).
With every technological advance come rumors about the end of the translator's profession. However, contrary to predictions, human work continues to be indispensable. This is because machine translation systems are able to provide a draft but not a final text that is coherent and concise. Therefore, the translator acts as a validator of the translation, a technical and terminology consultant, and also helps improve the automatic system. Additionally, literary texts, which require more translational freedom, and legal texts, which require more rigid interpretation, are types of text that cannot be accommodated by the automated system. Technology should not be considered a substitute for human work, but rather a tool to help expedite and facilitate translations.
|
